SEARCH     advanced search my account e-alerts subscribe register
Journal home
Advance online publication
Current issue
Press releases
For authors
For referees
Contact editorial office
About the journal
For librarians
Contact NPG
Customer services
Site features
NPG Subject areas
Access material from all our publications in your subject area:
Biotechnology Biotechnology
Cancer Cancer
Chemistry Chemistry
Dentistry Dentistry
Development Development
Drug Discovery Drug Discovery
Earth Sciences Earth Sciences
Evolution & Ecology Evolution & Ecology
Genetics Genetics
Immunology Immunology
Materials Materials Science
Medical Research Medical Research
Microbiology Microbiology
Molecular Cell Biology Molecular Cell Biology
Neuroscience Neuroscience
Pharmacology Pharmacology
Physics Physics
Browse all publications
21 October 1999, Volume 18, Number 43, Pages 5850-5859
Table of contents    Previous  Article  Next   [PDF]
Mouse ULK2, a novel member of the UNC-51-like protein kinases: unique features of functional domains
Jin Yan1, Hidehito Kuroyanagi2,3, Takuya Tomemori2, Noriko Okazaki1, Kuroiwa Asato4, Yo-ichi Matsuda4, Yo-ichi Suzuki2, Yasumi Ohshima5, Shohei Mitani6, Yasuhiko Masuho1, Takuji Shirasawa2,7 and Masa-aki Muramatsu1

1Helix Research Institute, 1532-3 Yana, Kisarazu, Chiba, 292-0812, Japan

2Department of Molecular Genetics, Tokyo Metropolitan Institute of Gerontology, 35-2, Sakaecho, Itabashi-ku Tokyo, 173-0015, Japan

3Laboratory of Cellular Biosynthesis, Institute of Molecular Bioscience, University of Tokyo, 1-1-1 Yayoui, Bunkyou-ku, Tokyo, 113-0032, Japan

4Laboratory of Animal Genetics, School of Agricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-01, Japan

5Department of Biology, Faculty of Science, Kyushu University, Fukuoka, 812-81, Japan

6Department of Physiology, Tokyo Women's Medical University School of Medicine, 8-1, Kawada-cho, Shinjuku-ku, Tokyo, 162-8666, Japan

7CREST, Japan Science and Technology Corporation, Japan

Correspondence to: Masa-aki Muramatsu, Helix Research Institute, 1532-3 Yana, Kisarazu, Chiba, 292-0812, Japan


The UNC-51 serine/threonine kinase of C. elegans plays an essential role in axonal elongation, and unc-51 mutants exhibit uncoordinated movements. We have previously identified mouse and human cDNAs encoding UNC-51-like kinase (ULK1). Here we report the identification and characterization of the second murine member of this kinase family, ULK2. Mouse ULK2 cDNA encodes a putative polypeptide of 1033 aa which has an overall 52% and 33% amino acid identity to ULK1 and UNC-51, respectively. ULKs and UNC-51 share a typical domain structure of an amino-terminal kinase domain, a central proline/serine rich (PS) domain, and a carboxy-terminal (C) domain. Northern blot analysis showed that ULK2 mRNA is widely expressed in adult tissues. In situ hybridization analysis indicated that ULK2 mRNA is ubiquitously localized in premature as well as mature neurons in developing nervous system. ULK2 gene was mapped to mouse chromosome 11B1.3 and rat chromosome 10q23 by FISH. HA-tagged ULK2 expressed in COS7 cells had an apparent molecular size of ~150 kDa and was autophosphorylated in vitro. Truncation mutants suggested that the autophosphorylation occurs in the PS domain. Although expression of ULK2 failed to rescue unc-51 mutant of C. elegans, a series of ULK2/UNC-51 chimeric kinases revealed that function of the kinase and PS domains are conserved among species, while the C domain acts in a species-specific manner. These results suggest that ULK2 is involved in a previously uncharacterized signaling pathway in mammalian cells.


protein kinase; signal transduction; UNC-51-like kinases; C. elegans; autophosphorylation


Protein kinases have been known to play crucial roles in nearly all aspects of cellular regulation, including cell growth, differentiation, and apoptosis in eukary-otes. Many protein kinases with specific or multiple functions have been characterized, which are diversed in primary structure, regulation modes and substrate specificities (Hunter, 1987; Cohen, 1992). An increasing number of protein kinase genes is being identified in the course of genome sequencing projects of various eukaryotes, and to date protein kinases constitute one of the largest gene superfamily (Tatusov et al., 1997). For example, the complete genome of S. cerevisiae and C. elegans have been surveyed and ~118 and ~435 protein kinase genes were found, respectively, both of which correspond to about 2% of the total genes (Chervitz et al., 1998). Phylogenetic analysis of protein kinases allows subgrouping of structurally related members, which also reflects related function or regulation (Hanks and Hunter, 1995). Of interest, most of the kinases found in yeast and nematode also have homologous counterparts in mammals and other species to constitute subfamilies.

We have previously cloned a mouse and a human cDNA encoding UNC-51-like kinase (ULK1) and determined its structure (Yan et al., 1998; Kuroyanagi et al., 1998). ULK1 is structurally related to protein kinases encoded by UNC-51 of C. elegans (Ogura et al., 1994), and with less homology to APG1 of S. cerevisiae (Matsuura et al., 1997). UNC-51 is a serine/threonine kinase essential for axonal elongation in C. elegans. Mutation of the UNC-51 gene results in paralyzed, egg-laying defective and dumpy phenotypes (Hedgecock et al., 1985). In such mutants, most of the neurons exhibit an aberrant axon elongation with premature termination or dysregulated numbers of axons in abnormal directions (Desai et al., 1988). Fine structural analysis indicates atypical membranous vesicles and cisternae-like structures in axons and ectopic varicosities near the cell bodies (McIntire et al., 1992).

On the other hand, S. cerevisiae APG1 was initially identified as one of the essential genes for autophagocytosis in yeast (Tsukada and Ohsumi, 1993), and was later shown to encode a serine/threonine kinase structurally related to UNC-51 (Matsuura et al., 1997). Autophagy in yeast is a protein degradation process induced by depletion of nutrients to balance energy consumption, and APG1 deficient yeast is lethal upon starvation (Takeshige et al., 1992). Structural similarity of UNC-51 and APG1 kinases and their involvement in distinct phenotypes in each organism suggest that they might be involved in a common fundamental pathway, such as regulation of membrane organization or movement of organelle (Matsuura et al., 1997). However, the molecular mechanism by which these kinases exert their physiological roles remains to be elucidated.

In the present study we report the cloning and characterization of a novel member of UNC-51-like kinase, ULK2. To study whether ULK2 shares common function with UNC-51, ULK2 was expressed in a mutant worm of unc-51. Although the original ULK2 did not rescue the phenotype, a series of chimeric kinases between ULK2 and UNC-51 revealed unique structural requirements for its function in C. elegans. Taken together with these results, we propose that ULK2 might be involved in regulation of elemental biological processes conserved among wide variety of species.


Cloning of mouse ULK2

In order to search other members of UNK-like kinases in mammals, the nucleotide and protein sequences of ULK1 were subjected to homology search against GenBank database. A mouse EST (GenBank accession number W29537) which has a high similarity to a portion of the coding region of ULK1 was found. To obtain the 5'- and 3'-sides of the transcript, oligonucleotide primers were designed according to this sequence as described in Methods, and nested PCR was applied using a mouse embryo cDNA library as a template. PCR using the 3'-specific primers generated four DNA fragments of 3.2, 4.0, 4.4 and 4.9 kb in size. Subcloning and sequence analysis revealed that these four cDNA clones contained a long overlapping sequence with only differences in the 3'-termini, indicating that all of these transcripts are derived from a single gene by alternative transcriptional termination. The 5'-end of the clone was also obtained by nested PCR using 5'-specific primers, and the consensus sequence of the entire cDNA was deduced from these sequences. The longest assembled cDNA was 5275 bp which contained an ORF of 1037 amino acids (Figure 1). The sequence around the first ATG codon at nt 42 - 44 was compatible with the Kozak's consensus (Kozak, 1986). The 3'-terminus of the longest cDNA ended with a poly(A)+ tail preceded by a polyadenylation signal at nt 5260 - 5266. Three shorter cDNA clones as just described had poly(A)+ attached at nt 3429, nt 4362, and nt 4728, probably utilizing the polyadenylation signals at nt 3414 - 3419, 4342 - 4347 and 4708 - 4713, respectively (underlined in Figure 1). The predicted protein sequence deduced from the ORF revealed that it had an overall sequence similarity with UNC-51 and ULK1. We therefore designated this clone as mouse ULK2 (deposited in GenBank, EMBL, and DDBJ, under the accession number of AB019577).

Structural comparison of ULK2, ULK1 and UNC-51

The protein sequences of mouse ULK2, ULK1 and C. elegans UNC-51 are aligned in Figure 2. The overall amino acid identity of ULK2 with ULK1 and UNC-51 was 52% and 33%, respectively. The entire molecule delineates three domains; the amino-terminal kinase domain, the central proline/serine rich (PS) domain and the carboxy-terminal conserved (C) domain.

The kinase domain of ULK2 possesses all of the 12 subdomain motifs recognized as being nearly invariant throughout the protein kinase superfamily (Hanks and Hunter, 1995). The sequence from the subdomain VIb and subdomain IX of ULK2 is characteristic of a serine/threonine kinase (Hanks and Hunter, 1995). The amino acid identity of ULK2 (1-271) with those of ULK1 (1 - 278) and UNC-51 (1 - 275) was 76% and 63%, respectively. The kinase domain of ULK2 showed lower homologies (30 - 33% amino acid identity) to other mammalian protein kinases such as PKCmu (GenBank, accession number Q15139), NRK2 (P51957), NEK1 (P51954) and, RSKIIalpha1 (P18653), all of which apparently belong to different serine/threonine kinase subfamilies.

The PS domain of ULK2 contains 12% of serine and 16% of proline residues. ULK2 (272 - 802) shares 38% and 20% identity with ULK1 (279 - 819) and UNC-51 (276 - 638), respectively. Although the PS domain is the least homologous domain, comparison with those of ULK2 and ULK1 reveals conserved short amino acid stretches, such as KKS(C/P)PVPVP, ASPPSL, CDTDDFV(M/L)VP, PIPVPTQV(R/H)NYQRIEQNL, RPY(S/T)PSP(L/Q)VGTIPE, GSPTK, GSTDSL, FTVGSP, and SVGSSSS (see Figure 2). Homology search of these sequences failed to identify similarities to other functionally defined motifs, and thus their roles are unknown at present.

The C domain of ULK2 (803 - 1037) and those of ULK1 (820 - 1051) and UNC-51 (639 - 856) shares 56% and 26% identity, respectively. When the C domain of UNC-51 (639 - 856) was used as a query to search for homologous sequence using BLAST, the most homologous mammalian sequences were those derived from the C domains of ULK1 and ULK2 (not shown). This reinforces the notion that ULKs and UNC-51 are in the same subfamily of serine/threonine kinase, and that ULKs may be a canonical UNC-51 homologue in mammals. The UNC-51 mutations attributing to unc-51 phenotype have often been found in this region (Ogura et al., 1994). Also UNC-14, an UNC-51 binding protein, interact with a region including this C domain (Ogura et al., 1997). These lines of evidence together with the relatively high sequence conservation point to the functional importance of the C domain.

Tissue distribution of ULK2 mRNA

We examined the tissue distribution of ULK2 mRNA by Northern blot analysis (Figure 3). ULK2 mRNA was detected as multiple bands distributing from 3.5 - 5.5 kb in heart, brain, spleen, lung, liver skeletal muscle, kidney and testis at varying degrees, indicating that ULK2 gene is ubiquitously expressed in various adult tissues. The longest transcript was most abundant in skeletal muscle, while the shortest transcript was most abundant in testis. Multiple bands seen in Northern blots are consistent with the existence of different polyadenylation sites as shown in cloned ULK2 cDNAs (see Figure 1). ULK1 transcript was detected as a single transcript of 4.7 kb in the same tissues examined (Yan et al., 1998). Therefore, both ULK1 and ULK2 mRNAs are ubiquitously expressed and failed to show a specific expression pattern.

ULK2 mRNA localization in mouse embryo

In order to analyse the spatial gene expression of ULK2 in mouse tissues, we performed in situ hybridization for mouse embryos. In the sagital section of a day 16 p.c. embryo, the transcript of ULK2 was detected in the developing brain and spinal cord (Figure 4a). In the embryonic brain, the antisense riboprobe detected in situ signals in developing cerebral cortex, thalamus, striatum, midbrain, and spinal cord (Figure 4a). In the cerebral cortex, ULK2 mRNA was ubiquitously transcribed in the neural cells of ventricular, subventricular and intermediate zones as well as in cortical plate (Figure 4c). In high power field view of the lower border of intermediate zone, in situ signals were localized to the differentiating immature neurons (Figure 4d). In the cortical plate, in situ grains were also localized to the maturing neurons (Figure 4a), suggesting that ULK2 is actively transcribed in maturing neural cells, irrespective of differentiation or maturation stages. The in situ signals were also localized to the differentiating and maturing neural cells in developing thalamus (Figure 4e) and spinal cord (data not shown) of a day 16 p.c. embryo. The signals detected by antisense probe were further confirmed by another riboprobe derived from a distinct part of ULK2 cDNA (data not shown), which were neither detected by sense probe (Figure 4b). This pattern of ULK2 expression is also shown in later stage of gestation (18 p.c.) as well as neonate mice (data not shown). These observations suggest a basic biological role of ULK2 in developing nervous system.

Expression and autophosphorylation of ULK2

We then examined the in vitro kinase activity of ULK2. To do this, a kinase negative mutant ULK2Delta was constructed as a control by replacing the conserved ATP-binding Lys with Thr. pHA - ULK2 and pHA - ULK2Delta plasmids which express HA-tagged proteins were transfected into COS7 cells, and cell lysates were subjected to immunoblotting analysis and immunoprecipitation kinase assay (Figure 5). Immunoblotting with an HA antibody identified a molecule of ~150 kDa which corresponds to HA - ULK2Delta (lane 1) and HA - ULK2 (lane 2) in cell lysates. The same lysates were subjected to the phosphorylation assay and a band corresponding to phosphorylated HA - ULK2 (lane 4) but not of HA - ULK2Delta (lane 3) was detected. This indicates that HA - ULK2 is phosphorylated and the phosphorylation is primarily due to autophosphorylation. To further determine the site of phosphorylation, expression plasmids for truncated ULK2 mutants, pHA - ULK2(793), pHA - ULK2(365), and pHA - ULK2(271) were transfected into COS7 cells and the same experiment was performed. Among the expressed ULK2 constructs (lanes 5 - 8), bands corresponding to phosphorylation of HA - ULK2 (lane 12), HA - ULK2(793) (lane 11), HA - ULK2(365) (lane 10), were readily detected. In contrast, phosphorylation of HA - ULK2(271) which encodes only the kinase domain was not detected (lane 9). These results indicate that ULK2 is autophosphorylated in the kinase domain or between 272 - 365 of the PS domain. Deletions of the domains up to ULK2(365) did not significantly change the level of autophos-phorylation. This suggests that the deleted domains may not be involved in regulating kinase activity, at least in vitro.

Complementation of C. elegans unc-51 mutants with chimeric ULK2 constructs

In order to analyse the functional conservation between ULK2 and UNC-51, we tested whether ULK2 expression can rescue the phenotypes of unc-51 mutants; paralyzed, egg-laying defective and dumpy. The ULK2/UNC-51 constructs used for the injection and the results are summarized in Figure 6. These constructs were placed under the control of the unc-51 promoter, and were coinjected with pEGFP, an EGFP-expressing plasmid under the same unc-51 promoter. This allowed us to monitor the expression of transgenes and also the morphology of neurons that express EGFP. Co-expression of EGFP and UNC-51 in unc-51 (e369) mutant generated phenotypically reverted worms, confirming that the expression of UNC-51 (pU51) could potentially rescue the phenotypes of unc-51 mutants as expected (Table 1). However, the extent of rescue activities depended on the pattern and level for the expression of transgenes as monitored by EGFP (data not shown). In contrast to UNC-51, the expression of ULK2 (pULK2f) failed to rescue any of the unc-51 phenotypes (Table 1).

We then examined ULK2/UNC-51 chimeric constructs to determine functional domains needed to rescue the unc-51 mutant phenotypes. As the kinase domains of UNC-51 and ULK2 show the highest sequence similarity, functional redundancy of the kinase between mouse and C. elegans was first examined. To do this, the kinase domain of UNC-51 was replaced with those of ULK2, ULK2Delta, and PKA which is a typical member of the serine/threonine kinase family. Expression of ULK2K/U51 could rescue the morphological and behavioral defects of unc-51 mutants to an equal extent observed for the transgene of UNC-51 (Table 1). However, expression of ULK2KDelta/U51, a kinase defective version resulted in a complete loss of rescuing capability (Table 1). In addition, expression of PKAK/U51, a chimera with the kinase domain of PKA, totally failed to rescue the mutant (Table 1). These findings together indicated that the kinase domains of UNC-51 subfamily share a biological function distinct from those of other kinases, which is attributable to their kinase activity.

Next we examined the functional redundancy of the PS and C domains. Despite a very low sequence similarity in the PS domains between mouse and C. elegans, expression of ULK2PS/U51 rescued the unc-51 phenotypes to an equal extent with those of UNC-51 and ULK2K/U51 (Table 1). These findings indicated that the PS domains of ULK2 and UNC-51 may share the same functions, while the C domain of C. elegans UNC-51 cannot be exchanged and are needed for the rescue of the unc-51 mutant phenotypes.

Chromosomal location of mouse ULK2

The chromosomal assignment of the ULK2 gene to mouse and rat chromosomes was made by direct R-banding FISH using mouse cDNAs as probes. The ULK2 gene was localized to mouse chromosome 11B1.3 and rat chromosome 10q23 (Figure 7). This was mapped in the region where the conserved linkage homology has been identified between the species (Yamada et al., 1994). In the region where ULK2 was mapped, ds (Ames dwarf), jvs (juvenile visceral steatosis), llgl (lethal giant lavae), vt (vestigial tail), sh2 (shaker 2), vb (vibrator), and ti (tipsy) were also mapped (Watkins-Chow et al., 1996).


We have previously reported cloning and characterization of the first mammalian UNC-51-like kinase, ULK1 from mouse and human (Yan et al., 1998; Kuroyanagi et al., 1998). In the present study, we identified and characterized a second murine member of the UNC-51-like kinase family, ULK2. Taking the evolutional distance between mammals and nematodes into consideration, conservation in the kinase domains of ULKs and UNC-51 (60 - 65% amino acid identity) is comparable to those in the other well characterized kinase subfamilies. For example, kinase domains of nematode and mammalian PKA (Gross et al., 1990), PKCbeta (Land et al., 1994), PKCdelta, (Tabuse et al., 1989), PAK (Chen et al., 1996) and PKB (Paradis and Ruvkun, 1998), have 80%, 76%, 64%, 49%, and 35%, amino acid identity, respectively. Apart from the similarity within the kinase domain, ULKs and UNC-51 retain evolutionary conserved domain structures consisting of the kinase, PS and C domains. As a yeast serine/threonine kinase APG1 is also homologous to UNC-51 (Matsuura et al., 1997) and ULK1 (Yan et al., 1998; Kuroyanagi et al., 1998), ULK2 also have an overall identity of 22% to APG1. Therefore, ULKs, UNC-51 and APG1 collectively constitute a novel subfamily of protein kinases. ULK1 and ULK2 may have originated by gene duplication during the course of evolution, since there is no other UNC-51-like kinase identified at present in the genome sequence of C. elegans or S. cerevisiae.

During preparation of the manuscript, an EST named KIAA0623 (GenBank, accession number AB014523) which encodes a putative human protein kinase of 1036 aa was reported (Ishikawa et al., 1998). Since human and mouse ULK1 proteins have an overall 90% identity, and mouse ULK2 protein and the ORF of human KIAA0623 have an overall 93% identity, we postulate that this represents the human ULK2 cDNA.

The kinase activity of ULK2 was proven by demonstrating its autophosphorylation in vitro. Truncation mutants demonstrated that autophosphorylation occurs in ULK2 (365) but not in ULK2 (271) which contains the cannonical kinase domain. Although we could not exclude the possibility that ULK2 (271) is defective for kinase activity by substantial deletion, we prefer to think that ULK2 is predominantly phosphorylated in the 271 - 365 region of the PS domain. In any case, this does not exclude the possibility that there are other phosphorylations. ULK1 was also autophosphorylated and its truncation gave a similar result (Yan et al., 1998), suggesting that autophosphorylation in the PS domain might be a common biochemical character for this kinase subfamily. Since there are many serine residues in this domain, ULKs might be phosphorylated in multiple sites. The exact phosphorylated residue(s) and the consequence of the phosphorylation needs to be determined.

To obtain clues for cellular function of ULK2, we employed complementation studies in an unc-51 mutant of C. elegans. The kinase domains of ULK2 and UNC-51 were exchangeable in this study, showing that they can exert similar effects. Given that both kinase domains are 63% identical, it is conceivable that they could recognize and phosphorylate the same physiological substrate(s) in C. elegans. Further studies revealed functional conservation and difference of the PS and C domains between different species. The PS domain of ULK2, in the context of a chimeric kinase could also operate in C. elegans. Since PS domain is least homologous and still functionally redundant, this domain may serve as a spacer region. Alternatively, the PS domain might be needed for the acceptor sites for phosphorylation to render charge effects.

More importantly, exchange of the C domain of ULK2 to that of UNC-51 was sufficient to rescue the unc-51 phenotype. This indicates that the C domain has an essential function which acts in a species-specific manner. Although the primary structure of the C domain is relatively well conserved, it appears that the C domain has evolutionarily diverged so that it cannot operate between different species. The critical role of the C domain of UNC-51 in C. elegans revealed by our study is consistent with two lines of evidence. First, many of the unc-51 mutants have missense or nonsense mutations which leads to deletion or mutation of the C domain (Ogura et al., 1994). Second, UNC-14, a UNC-51 binding protein that is proposed to be a regulator interacts with the carboxy-half of UNC-51 including the C domain (Ogura et al., 1997). According to this scheme, it is tempting to speculate that the chimeric kinase with the C domain of UNC-51 could bind to UNC-14 and elicit function, while the C domain of ULK2 could not do so. It would be interesting to study in the future whether there is a mammalian counterpart of UNC-14. The precise role of C domain still remains to be clarified. Since deletion of the C domain did not affect in vitro autophosphorylation activity of ULK2 and ULK1 (Yan et al., 1998), it is unlikely to be a regulatory domain for kinase activity, but may have other role(s) such as an adapter domain to recruit the kinase to an appropriate physiological environment.

Dysfunction of UNC-51 in C. elegans leads to atypical membranous vesicles and cisternae-like structures in axons, and dysfunction of APG1 interferes the autophagic process. Taking the structural similarities and mutant phenotypes into account, APG1 and UNC-51 might be involved in elemental biochemical steps, such as regulation of membrane organization or movement of organelle (Matsuura et al., 1997). Whether or not ULKs are involved in analogous biological processes in mammalian cells still needs to be elucidated. The studies using recombinant ULKs to clarify their mode of regulation, substrate specificity as well as their physiological function are underway.

Materials and methods

Cloning of ULK2 cDNA

To obtain full length ULK2 cDNA, 5' and 3' nested PCR was performed using pSPORT mouse brain library (GIBCO BRL, Gaithersburg, MD, USA) and LA - PCR kit (TAKARA, Kyoto, Japan). The 5'-specific PCR primers for the pSPORT vector were A1: 5'-cac aca gga aac agc tat gac cac tag-3' and A2: 5'-aag cta ttt agg tga cac tat aga agg tac-3'. The 3'-specific PCR primer for the pSPORT vector was A3: 5'-gac tca cta tag gga gag agc tat gac gtc-3'. The ULK2 gene specific PCR primers designed from EST (W29537) were A4: 5'-gag acc tgg cag att att tgc aag cta aag-3', A5: 5-ctc tga gtg aag ata cta tca gag tgt ttc-3', A6: 5'-aaa gac aga gtt ggg caa ttc ctg aac atc-3', and A7: 5'-gag cgc tac gat gtt ttc atg ctg aag ctc-3'. To obtain 3' portion of ULK2 cDNA, the first PCR reaction was done using A3 and A4 primers and pSPORT mouse brain library as a template, and the second PCR reaction was done by using A3 and A5 primers, and 1/50 of first PCR reaction mixture as a template. To obtain 5' portion of ULK2 cDNA, first PCR reaction was done using A1 and A6 as primers and pSPORT mouse brain library as a template and the second PCR reaction was done by using A2 and A7 primers, and 1/50 of first PCR reaction product as a template. The amplified bands were subcloned into pT7-7 vector (Novagene, Madison WI, USA) and sequenced by an ABI377 (Perkin-Elmer, Foster City, CA, USA) sequencer.

Construction of ULK2 plasmids

ULK2 cDNA was inserted into XhoI - NotI cloning sites of a pME18S-3HA vector (Yan et al., 1998), to construct pHA - ULK2. A plasmid pHA - ULK2Delta which express a kinase defective mutant with the ATP-binding Lys substituted to Thr, was constructed by using a primer 5'-gtg gct att act agt att aat-3', and a double-stranded site directed mutagenesis kit (Stratagene, La Jolla, CA, USA). Expression plasmids for truncation mutants pHA - ULK2 (271), pHA - ULK2 (365), and pHA - ULK2 (793) were constructed by using primers, 5'-cct ttc ctt tag caa gtt cca-3', 5'-gac cac tca tag gac atg cca-3', and 5'-cta aga tac tag cct tat ggt-3', respectively. Mutants were checked by sequencing and sublconed into the pME18S-3HA vector. Chimeric kinases of ULK2 and UNC-51 were generated as follows. A 4 kb PstI - BstXI fragment from unc-51 genome DNA clone, pF14 (Ogura et al., 1994) was subcloned into PstI - BamHI site of pPD49.26 to generate an unc-51 promoter vector, pU51P. pEGFP was generated by subcloning 0.8 kb SalI - NotI fragment of pEGFP-N2 (Clontech, Palo Alto, CA, USA) into SalI - SacI restricted pU51P. pU51 was generated by subcloning 2.9 kb SalI - SpeI fragment from UNC-51 cDNA clone, pBLO (Ogura et al., 1994) into SalI - SacI site of pU51P. pULK2DeltaK/U51 and pPKAK/U51 were generated by replacing BstXI - XhoI fragment of pU51 with PCR-amplified ULK2, ULK2Delta or PKA cDNA, respectively. pULK2PS/U51 was generated by replacing 1.8 kb BamHI fragment of pULK2K/U51 with PCR-amplified ULK1 and UNC-51 cDNA fragments. pULK2f was generated by replacing 2.3 kb BamHI - SacI fragment of pULK2K/U51 with PCR-amplified ULK2 cDNA.

Transfection and immunoblotting analysis of ULK2

Expression plasmids for HA-tagged ULK2 constructs were transfected into COS7 cells using Lipofectamine (GIBCO - BRL, Gaithersburg, MD, USA), according to manufacturer's protocol. Cells were harvested 48 h after transfection, and cell lysates were prepared in an extraction buffer containing 50 mM Tris HCl pH 7.5, NaCl 150 mM, Triton-X100 0.1%, and PMSF 1 mM. Samples were fractionated by 7.5% SDS - PAGE and transferred to PDVF membrane (BIO-RAD, Hercules, CA, USA). After blocking with 5% nonfat milk in T-TBS (TBS with 0.1% Tween 20), immunodetection was done by using 12CA5 monoclonal anti-HA antibody (Boehringer Mannheim, Mannheim, Germany) as a first antibody and HRP-conjugated anti-mouse Ig antibody (Amersham, Buckinghamshire, UK) as a second antibody, followed by visualization with ECL detection kit (Amersham, Buckinghamshire, UK).

In vitro kinase assay

COS7 cell lysates expressing HA-tagged ULK2 or mutants were treated with 10 mug of 12CA5 and then with 20 mul of protein G-Sepharose (Pharmacia, Uppsala, Sweden). Immune complex was collected by centrifugation, washed twice in the extraction buffer and one in a kinase buffer (50 mM HEPES pH 7.5, 10 mM Mg(OAc)2 and 1 mM DTT). The immune complex was incubated for 15 min at 30°C in a kinase buffer containing 10 muCi of [gamma-32P]ATP (>3000 Ci/mmol) in a total volume of 25 mul. The reaction was terminated by the addition of Laemmli sample buffer, separated by 7.5% SDS - PAGE, and analysed by BAS-2000 Bio-image analyser (Fuji Photo Film, Tokyo Japan).

Northern blotting

A 0.75 kb BgII fragment from ULK2 cDNA was labeled using Megaprime kit (Amersham, Buckinghamshire, UK) and hybridized to a multiple tissue Northern blot membrane (Clontech, Palo Alto, CA, USA) which contains 2 mug of mRNA from various tissues. The blots were hybridized at 65°C for 2 h in QuikHyb (Stratagene, La Jolla, CA, USA) and washed twice with 2´SSC, 0.1% SDS for 20 min each at room temperature, followed by a final wash with 0.2´SSC, 0.1% SDS for 20 min at 65°C. The blots were analysed by BAS-2000 Bio-image analyser.

In situ hybridization

Sections were prepared from C57BL/6 mouse embryos at day 16 as described previously (Shirasawa et al., 1993). For the antisense riboprobe of mouse ULK2, pGEM7Zf(+) plasmid (Promega, Madison, WI, USA) harboring 449 bp EcoRI fragment (nt 242 - 670) from mouse ULK2 cDNA was linealized with XbaI and transcribed with SP6 polymerase according to manufacture's conditions (Promega) with [35S]-UTPalphaS (740 MBq/ml, Amersham). The sense riboprobe was linealized with SmaI and transcribed with T7 polymerase. Hybridizations were performed essentially as described (Shirasawa et al., 1993) with 5´107 c.p.m./ml radiolabeled probe. Sections were treated with 50 mug/ml RNase A at 37°C for 30 min, after the hybridization. Autoradiography was performed using NTB-3 emulsion (Kodak, New Haven, CT, USA). After 1 week exposure at 4°C, the slides were developed by Kodak D-19 and fixed by Fujifix. Sections were counterstained with hematoxylin and eosin (H&E).

Chromosome preparation and in situ hybridization

The direct R-banding FISH method was used for chromosomal assignment of the ULK2 gene to mouse chromosomes. Preparation of R-banded chromosomes and FISH were performed as previously described (Matsuda et al., 1992; Matsuda and Chapman, 1995). Mitogen-stimulated mouse and rat splenocyte culture was synchronized by thymidine block, and the incorporation of 5-bromodeoxyuridine during the late replication stage was made for differential replication staining after the release of excessive thymidine. R-band staining was performed by exposure of chromosome slides to UV light after staining with Hoechst 33258.

The chromosome slides were hardened at 65°C for 2 h and then denatured at 70°C in 70% formamide in 2´SSC and dehydrated in a 70 - 85 - 100% ethanol series at 4°C. The mouse ULK2 fragment inserted in pBluescript KS(-) was labeled by nick translation with biotin 16-dUTP (Boehringer Mannheim) following the manufacturer's protocol. The labeled DNA fragment was ethanol precipitated with salmon sperm DNA and E. coli tRNA, and then denatured at 75°C for 10 min in 100% formamide. The denatured probe was mixed with an equal volume of hybridization solution to make final concentration of 50% formamide, 2´SSC, 10% dextransulfate, and 2 mg/ml BSA (Sigma). 20 mul mixture containing 250 ng labeled DNA was put on the denatured slide, covered with parafilm and incubated overnight at 37°C. The slides were washed for 20 min in 50% formamide in 2´SSC at 37°C, and in 2´SSC and 1´SSC for 20 min each at room temperature. After rinsing in 4´SSC, they were incubated under coverslip with anti-biotin antibody (Vector Laboratories) at a 1 : 500 dilution for 1 h at 37°C. After washing with 4´SSC, 0.1% Nonident P-40 in 4´SSC, 4´SSC for 10 min on the shaker, the slides were rinsed with 2´SSC and stained with 0.75 mug/ml propidium iodide. Excitation at wave length 450 - 490 nm (Nikon filter set B-2A) and near 365 nm (UV-2A) were used for observation. Kodak Ektachrome ASA100 films were used for microphotography.

Generation of transgenic worms

The unc-51 mutant (e369) was provided by Caernorhabditis Genetics Center (USA). Worms were grown on NGM agar plates at 20°C fed with the OP50-1 strain of E. coli (Way and Chalfie 1988). A 9 : 1 mixture of pEGFP and other rescue constructs was microinjected into the syncytial gonads in adult hermaphrodites of unc-51 (e369) mutants at a total concentration of 200 ng/mul following the method described previously (Mello et al., 1991). Animals with EGFP expression in the neurons were selected as transgenic worms. The expression pattern of EGFP was apparently the same as previously observed (Ogura et al., 1994).


J Yan and H Kuroyanagi made equal contributions in this study. We thank Y Fujita for the photographical assistance.


Chen W, Chen S, Yap SF and Lim L. (1996). J. Biol. Chem. 271, 26362-26368. Article MEDLINE

Chervitz SA, Aravind L, Sherlock G, Ball CA, Koonin EV, Dwight SS, Harris MA, Dolinski K, Mohr S, Smith T, Weng S, Cherry JM and Botstein D. (1998). Science 282, 2022-2028. Article MEDLINE

Cohen P. (1992). Trends Biochem. Sci. 17, 408-413. MEDLINE

Desai C, Garriga G, McIntire SL and Horvitz HR. (1988). Nature 336, 638-646. MEDLINE

Gross RE, Bagchi S, Lu X and Rubin CS. (1990). J. Biol. Chem. 265, 6896-6907. MEDLINE

Hanks SK and Hunter T. (1995). FASEB J. 9, 579.

Hedgecock EM, Culotti JG, Thomson JN and Perkins LA. (1985). Dev. Biol. 111, 158-170. MEDLINE

Hunter T. (1987). Cell 50, 823-829. MEDLINE

Ishikawa K, Nagase T, Suyama M, Miyajima N, Tanaka A, Kotani H, Nomura N and Ohara O. (1998). DNA Res. 5, 169-176. MEDLINE

Kozak M. (1986). Cell 44, 283-292. MEDLINE

Kuroyanagi H, Yan J, Seki N, Yamanouchi Y, Suzuki T, Takano T, Muramatsu M and Shirasawa T. (1998). Genomics 51, 76-85. MEDLINE

Land M, Islas-Trejo A, Freedman JH and Rubin CS. (1994). J. Biol. Chem. 269, 9234-9244. MEDLINE

Matsuda Y, Harada Y, Natsume-Sakai S, Lee K, Shiomi T and Chapman VM. (1992). Cytogenet. Cell Genet. 61, 282-285. MEDLINE

Matsuda Y and Chapman VM. (1995). Electrophoresis 16, 261-272. MEDLINE

Matsuura A, Tsukada M, Wada Y and Ohsumi Y. (1997). Gene 192, 245-250. Article MEDLINE

McIntire SL, Garriga G, White J, Jacobson D and Horvitz HR. (1992). Neuron 8, 307-322. MEDLINE

Mello CC, Kramer JM, Stinchcomb D and Ambros V. (1991). EMBO J. 10, 3959-3970. MEDLINE

Ogura K, Wicky C, Magnenat L, Tobler H, Mori I, Muller F and Ohshima Y. (1994). Genes Dev. 8, 2389-2400. MEDLINE

Ogura K, Shirakawa M, Barnes TM, Hekimi S and Ohshima Y. (1997). Genes Dev. 11, 1801-1811. MEDLINE

Paradis S and Ruvkun G. (1998). Genes Dev. 12, 2488-2498. Article MEDLINE

Shirasawa T, Akashi T, Sakamoto K, Takahashi H, Maruyama N and Hirokawa K. (1993). Dev. Dyn. 198, 1-13. MEDLINE

Tabuse Y, Nishiwaki K and Miwa J. (1989). Science 243, 1713-1716. MEDLINE

Tatusov RL, Koonin EV and Lipman DJ. (1997). Science 278, 631-637. Article MEDLINE

Takeshige K, Baba M, Tsuboi S, Noda T and Ohsumi Y. (1992). J. Cell. Biol. 119, 301-311. MEDLINE

Tsukada M and Ohsmumi Y. (1993). FEBS Lett. 333, 169-174. MEDLINE

Watkins-Chow D, Roller M, Newhouse MM, Buchberg AM and Champer SA. (1996). Mammal. Genome 6, S201-S220.

Way JC and Chalfie M. (1988). Cell 54, 5-16. MEDLINE

Yamada J, Kuramoto T and Serizawa T. (1974). Mammal. Genome 5, 63-83.

Yan J, Kuroyanagi H, Kuroiwa A, Matsuda Y, Tokumitsu H, Tomoda T, Shirasawa T and Muramatsu M. (1998). Biochem. Biophys. Res. Comm. 246, 222-227. MEDLINE


Figure 1 Nucleotide sequence and deduced amino acid sequence of mouse ULK2 cDNA. The amino acid sequence is shown below the nucleotide sequence by a single-letter code. Four cloned cDNAs had different 3'-termination followed by the polyadenylation signals as underlined. An asterisk denotes the stop codon. The GenBank accession number for the nucleotide sequence is AB019577

Figure 2 Alignment of mouse ULK2, ULK1, and C. elegans UNC-51 proteins. mULK2, mULK1 and UNC-51 protein sequences were first aligned by ALIGN ver 2. and then modified by eyes. Gaps (-) were introduced to maximize the alignment. Conserved amino acid residues are shown in black boxes. Kinase domain structures (I - XI), and boundaries of PS and C domains are indicated below the alignment

Figure 3 Northern blot analysis of mouse ULK2 gene. A multiple tissue blot filter carrying 2 mug of poly(A)+RNA from various adult tissues was hybridized with the ULK2 cDNA probe. Hybridization bands distributing from 5.5 - 3.5 kb are detected at various levels in all tissue examined. The multiple bands may reflect the heterogeneity in the 3'-termination sites of ULK2 mRNA

Figure 4 In situ hybridization of ULK2 mRNA in mouse embryo (a and b) sagital sections of E16 mouse embryo hybridized with antisense (a) and sense (b) riboprobes for ULK2 in the dark field presentations. In situ signals were detected in the brain and spinal cord. (c - e) bright-field photomicrographs of developing cerebral cortex (c and d) and thalamus (e) hybridized with antisense riboprobe for ULK2 and counterstained with H&E. (c) ULK2 transcript was widely detected in ventricular (VZ), subventricular (SV), intermediate zones (IZ) as well as cortical plate (CP). Scale bar: 300 mum. (d) Higher magnification of the lower border of intermediate zone. Scale bar: 100 mum. (e) Higher magnification of developing thalamus. Scale bar: 100 mum

Figure 5 Expression and autophosphorylation of ULK2 and truncation mutants. Immunoblotting analysis of HA tagged-ULK2 constructs expressed in COS7 cells using an HA-antibody. Lane 1: ULK2Delta, lane 2: ULK2, lane 5: ULK2 (271), lane 6 ULK2 (365), lane 7: ULK2 (793), and lane 8: ULK2. In vitro kinase assay of HA-tagged ULK2 and mutants. Lane 3: ULK2Delta, lane 4: ULK2, lane 9: ULK2 (271), lane 10: ULK2 (365), lane 11: ULK2 (793), and lane 12: ULK2. The products are shown by asterisks

Figure 6 ULK2/UNC-51 chimeric constructs and complementation of unc-51 mutants. Schematic representations of ULK2 and UNC-51 chimeric protein kinases are shown together with their capability to rescue unc-51 mutants. The numbers indicate the position of amino acid sequences. An asterisk indicates the point mutation site in ULK2. (+) indicates that the construct is capable of rescuing paralyzed, egg-laying defective and dumpy phenotype, while (-) indicates that it is incapable of doing so. The detailed result of the experiment is shown in Table 1

Figure 7 Chromosome mapping of mouse ULK2 gene. Chromosomal localization of the ULK2 gene on mouse (a, b, c) and rat (d, e) R-banded chromosomes. The mouse ULK2 cDNA fragment was used as a biotinylated probe. The hybridization signals are indicated by arrows. The ULK2 gene was localized to mouse chromosome 11B1.3 and rat chromosome 10q23. The metaphase spreads were photographed with Nikon B-2A (a, c, d, e), and UV-2A (b) filters. R-band and G-band patterns are demonstrated in (a, c, d, e) and (b), respectively


 Complementation analyses of unc-51 (e369) mutants by ULK2/UNC-51 chimeric constructs

Received 25 January 1999; revised 11 May 1999; accepted 21 May 1999
21 October 1999, Volume 18, Number 43, Pages 5850-5859
Table of contents    Previous  Article  Next    [PDF]